iiinova — where your digital self keeps evolving

Reposted by adn 👣

Serge Belongie @serge.belongie.com · 08/10/2026

Registration is open for the @ellis.eu UnConference! 🎉 This is an outstanding community driven event for the machine learning community, taking place the day before @neuripsconf.bsky.social Paris 🗓️ ellis-unconference.github.io

ellis-unconference.github.io

ELLIS UnConference

132

Reposted by adn 👣

Yoshitomo Matsubara @yoshitomo-matsubara.net · 07/07/2026

Call for emergency reviewers #NeurIPS2026 I'm looking for two emergency reviewers: - Multimodal (text + vision) benchmark - AI for Science (text-focused) benchmark If you have experience in either domains, send me a DM including your OpenReview profile and Google Scholar links

055

Reposted by adn 👣

ELLIS @ellis.eu · 26/05/2026

CVPR@Paris 2026 🇫🇷 — June 1st, co-organised by ELLIS Unit Paris. A one-day local event ahead of CVPR, open to all. Oral & poster sessions for CVPR 2026, CVPR workshops & ICLR 2026 papers. 🔗 cvprinparis.github.io/CVPR2026InPar…

CVPR@Paris 2026, co-organised by ELLIS Unit Paris.

0244

Reposted by adn 👣

European Conference on Computer Vision #ECCV2026 @eccv.bsky.social · 21/05/2026

We are happy to share that 9 tutorials have been accepted for #ECCV2026. The tutorials will take place on the 8th and 9th September. The chosen list can be found here, with more information to be added. eccv.ecva.net/Conferences/...

011

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 29/04/2026

Parallels are drawn between Physics and the Theory of Deep Learning, I love it. arxiv.org/pdf/2604.21691

1203

Reposted by adn 👣

ELLIS @ellis.eu · 27/04/2026

Where learning theory meets modern ML: Learning Theory Summer School (22-25 Jun) + Workshop (26-27 Jun) 📍 Copenhagen 🇩🇰 📚 Generalisation, optimisation, RL & responsible ML By: @aicentre.dk · Danish Data Science Academy ⏰ Register by 25 May 🔗 All info: ltsscopenhagen.github.io

141

Reposted by adn 👣

Dima Damen @ECCV 2026 @dimadamen.bsky.social · 27/04/2026

From the V-Jepa2 presentation: Action anticipation in EPIC-KITCHENS is "a very challenging task and still there’s head room for improvement on this task". V-Jepa2 establishes a new SOTA but still <65% on both verb and noun acc. Mido Assran at the @iclr-conf.bsky.social workshop on World Models

0131

Reposted by adn 👣

European Conference on Computer Vision #ECCV2026 @eccv.bsky.social · 19/04/2026

Message from #ECCV2026 Program Chairs: We recognise that missing review warning emails may feel excessive, especially for those of you who consistently complete your reviews on time (thank you).

194

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 19/04/2026

What are your all time favorite textbooks? Here are a few of mine.

4194

Reposted by adn 👣

Yoshitomo Matsubara @yoshitomo-matsubara.net · 20/04/2026

Looking forward to upgrading all my Ubuntu machines to 26.04 LTS this week Lots of things to follow the upgrade

222

Reposted by adn 👣

European Conference on Computer Vision #ECCV2026 @eccv.bsky.social · 19/04/2026

In case you haven't seen the mails, we are now a little more than two days from the #ECCV2026 review deadline 😉 Please note the deadline is 11 pm CET, April 21st!

media.tenor.com

a black and white clock with a spiral in the center

ALT: a black and white clock with a spiral in the center

053

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 13/04/2026

That was not on my Bingo card arxiv.org/pdf/2604.06425 Also: x.com/yuntiandeng/...

2353

adn 👣 @adnrv.bsky.social · 08/04/2026

Whether the work is accepted with a WA seems to be tied to the number of accepted works, which shouldn't be imposed! It seems that @icmlconf.bsky.social is trending towards bad practices of capping acceptance rates and defining a capacity (a big no no) for these recommendations to hold WTF?!

000

adn 👣 @adnrv.bsky.social · 08/04/2026

If we allow the weak versions of the decisions, and now allow for weak decisions from the ACs, we just keep moving the decision process higher in the stack of abstractions of people that are less connected to the work

110

adn 👣 @adnrv.bsky.social · 08/04/2026

The job of the program committee of confs is to make a binary decision The decision should be based on the technical contributions, correctness and soundness of the work If the work has no more concerns, accept it!! It is binary not fuzzy -- #ICML weak accept frustration

110

Reposted by adn 👣

Yoshitomo Matsubara @yoshitomo-matsubara.net · 07/04/2026

As an AC of #ICML2026, I replied to all AC confidential comments at least for acknowledging their comments because that's how I want an AC to behave rather than keep silence As an author, I submitted AC confidential comments. There is no response or indication that they read 🫩

161

Reposted by adn 👣

ELLIS @ellis.eu · 06/04/2026

🏹 Job alert: Postdoctoral position in method development for spatio-temporal medical data at the Arctic University of Norway (UiT) 📍 Tromsø 🇳🇴 ⏰ Apply by 26 April 🔗 bit.ly/4skTQLe

bit.ly

Postdoctoral Research Fellow in Machine Learning (297935) | UiT The Arctic University of Norway

Job title: Postdoctoral Research Fellow in Machine Learning (297935), Employer: UiT The Arctic University of Norway, Deadline: Sunday, April 26, 2026

042

Reposted by adn 👣

ELLIS unit Jena @ellisunitjena.bsky.social · 25/03/2026

👨‍🔬Welcome to ELLIS Unit Jena, Prof. Dr. Alexander Binder ! Professor for Multimodal Machine Learning and Principal Investigator at @scadsai.bsky.social and @unileipzig.bsky.social focuses on one of the most critical challenges in modern AI: making multimodal models more explainable and trustworthy.

041

Reposted by adn 👣

Yoshitomo Matsubara @yoshitomo-matsubara.net · 26/03/2026

ICLR 2026: Identity leakage during the review process ICML 2026: Two-policy framework for "LLM reviewing" NeurIPS 2026: Sanction list(s?) 2026 must be a tough year for ML conference organizers...

1101

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 23/03/2026

Quo vadis deep learning?

1102

Reposted by adn 👣

NeurIPS Europe @neuripseurope.bsky.social · 23/03/2026

Today NeurIPS is announcing our official satellite event in Paris. After responding to the call from Ellis following the success of EurIPS in December, we are pleased to reach a new milestone by joining forces with the NeurIPS organizing committee for the 2026 edition.

18933

Reposted by adn 👣

ELLIS @ellis.eu · 23/03/2026

💬 AI for Good Webinar Series 'From Molecules to Models' - today with Alexander Binder 🇩🇪 on 'Explainable AI in cases when you cannot spot the dogs or cats'. 📅 Mon, Mar 23 🕓 17:00-18:00 CET 🔗 Registration: bit.ly/4sP4URk ℹ️ More info: bit.ly/47Qclim

ELLIS AI for Good Webinar Series 'From Molecules to Models' with Alexander Binder on 'Explainable AI in cases when you cannot spot the dogs or cats'.

082

Reposted by adn 👣

European Conference on Computer Vision #ECCV2026 @eccv.bsky.social · 13/03/2026

#ECCV2026 🇸🇪 received 10,665 submissions going to review. That is an ECCV record. +24% vs #ECCV2024 (8,585) #ECCV2022: 6,773

1174

Reposted by adn 👣

Eric Brachmann @ericbrachmann.bsky.social · 06/03/2026

"Authors should not use negative v-spaces to change the template layout." The template layout:

0285

adn 👣 @adnrv.bsky.social · 05/03/2026

before that you had the manual which is very good

000

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 04/03/2026

TIL that - "state of the art" is a noun, so we compare to the state of the art - "state-of-the-art" is an adjective, so we can write about the state-of-the-art model on, say, edge-detection or thresholding or binarization.

3193

Reposted by adn 👣

European Conference on Computer Vision #ECCV2026 @eccv.bsky.social · 27/02/2026

The #ECCV2026 paper registration deadline has passed. No new papers may be registered. The author list is final. No additions allowed; reordering is permitted before submission deadline. Please do not email us about these matters. The rules are strict, no exceptions.

173

adn 👣 @adnrv.bsky.social · 27/02/2026

@arxiv.bsky.social can we get compound last names working on the website ? Two last names get confused with names in the current setup We need a way to define what is a first name and a last name -- with regards all latin people

010

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 26/02/2026

Since 2 or 3 conference editions I now see authors withdrawing papers _after_ the decisions have been communicated. Several papers of my batch have been rejected and then later withdrawn by the authors. Why?

361

Reposted by adn 👣

Walter Scheirer @wjscheirer.bsky.social · 24/02/2026

The #WACV2026 main conference open access proceedings is online: openaccess.thecvf.com/WACV2026 Workshops are coming soon... enjoy Arizona everyone!

1157

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 23/02/2026

We didn't submit to CVPR this year but from folks around me I heard that it was brutal. Lots of rejections of papers I would have thought would have passed.

2151

Reposted by adn 👣

David Picard @davidpicard.eurosky.social · 23/02/2026

Hey @cvprconference.bsky.social Do you know when the opt-in for the findings will be available?

161

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 22/02/2026

I hate the ECCV template so much. Why can't it just die? Nobody prints proceedings into LNCS volumes anymore. My favorite templates: 1. IEEE/CVPR 2. NeurIPS ... ... 39,2312,321. Edging tikz commands in morse code onto the skin of a dead walrus ... ... Springer/ECCV

4475

Reposted by adn 👣

#CVPR2026 @cvprconference.bsky.social · 21/02/2026

Special shoutout to our #CVPR2026 technical chair @yoshitomo-matsubara.net for dealing with the technical logistics behind the review process and interfacing with OpenReview. This was an especially challenging job this round with the extraordinary wrinkles added by the OpenReview leak.

1142

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 18/02/2026

👍 These are scientific conferences, and scientific results should/must allow us to gain new insights and understanding. If authors want to brag about (I mean, "report") new capabilities, I think a blog post or LinkedIn post might be more appropriate.

251

Reposted by adn 👣

David Picard @davidpicard.eurosky.social · 18/02/2026

Full support on this statement. Many papers achieve state-of-the-art by using a random collection of tricks that are not properly investigated. If that's how performances are increased, at least ablate it thoroughly so that it's science and not alchemy.

371

Reposted by adn 👣

Kosta Derpanis @csprofkgd.bsky.social · 18/02/2026

Nothing controversial here. A paper should be considered holistically. Simply getting a state-of-the-art number does automatically mean paper acceptance.

341

Reposted by adn 👣

Transactions on Machine Learning Research @tmlrorg.bsky.social · 12/02/2026

TMLR's 2025 annual report! The word of the year is growth: whether it's our number of submissions, published papers, or reviewers, pretty much everything has increased ~50%. EXCEPT: our review period, which is still shorter than NeurICMLR (91 vs 126 days) Read on for more! 1/n

1239

Reposted by adn 👣

ICML Conference @icmlconf.bsky.social · 12/02/2026

ICML 2026 Ethics Violation Reporting can be done via the following (anonymous) form: docs.google.com/forms/d/e/1F... You can also email the PC chairs directly (from your own email or an anonymous one) at this addres: program-chairs@icml.cc

docs.google.com

ICML 2026: Ethics Violation Reporting

This form can be used to report potential violations of ICML's ethics guidelines and code of conduct, as outlined at https://icml.cc/Conferences/2026/CallForPapers. All contact information are option...

021

Reposted by adn 👣

ICML Conference @icmlconf.bsky.social · 12/02/2026

Machine learning is an *international community*. Every community is based on mutual trust and respect. But communities are not robust to betrayals of these principles. Protect our community by reporting, e.g., review collusion or submission deanonymization via this form 👇

1128

adn 👣 @adnrv.bsky.social · 09/02/2026

Undergrads and master students are not peers yet... @icmlconf.bsky.social filter more the pool of reviewers 🙏

000

Reposted by adn 👣

Eric Brachmann @ericbrachmann.bsky.social · 04/02/2026

It's time, dear #CVPR2026 reviewers. Time to decide.

1153

adn 👣 @adnrv.bsky.social · 05/02/2026

I was thinking something like a bank of Gabor filters but data-driven (at least I remember it was)

100

adn 👣 @adnrv.bsky.social · 05/02/2026

we are all old in our hearts... 🙈

010

adn 👣 @adnrv.bsky.social · 05/02/2026

I was thinking about a follow up on that, but I couldn't find it. Something like what @davidpicard.bsky.social showed in his reply. Maybe I'm misremembering it, but I think there was a paper about it.

100

adn 👣 @adnrv.bsky.social · 04/02/2026

I'm looking for a paper (pre deep learning) about learning a filter bank from images. It showed that a data-driven filter bank was possible. It resembled the filters learned by CNNs. Does anyone know what I am talking about? Some older people with pre deep learning memory?

311

Reposted by adn 👣

Christian Wolf @chriswolfvision.bsky.social · 02/02/2026

"Cherry picking" in scientific papers. (Repost of an older post on some other site in 2021).

1335